An experiment with a vertical slot with horizontally seeping water with a dye diffusing from below was performed to help validate and visualize the Q-equivalent model, which describes the mass transfer rate from a source into flowing water, such as that in a repository for nuclear waste. The Q-equivalent model is used for quantifying mass transport in geological repositories. However, the tracer propagated much slower and to a lesser extent than predicted by the model. It was found that the tracer gave rise to a small density gradient that induced buoyancy-driven flow, overwhelming that driven by the horizontal hydraulic gradient. This dramatically changed the mass transfer from the dye source into the water in the slot. For the release of contaminants, this can have detrimental as well as beneficial effects, depending on whether positive or negative buoyancy is induced. These observations led to an analysis of when and how density differences in a repository can influence the release and further fate of escaping radionuclides in waste repositories. This and other experiments also showed that laboratory experiments aimed at visualizing flow and mass transfer processes in fractures could be very sensitive to the heating of the dye tracers by the lighting in the laboratory.